Building Search Database Construction Using Image Segmentation and Keypoint Masking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Image-based building search services face ambiguity and errors due to multiple buildings being captured in a single scene or presence of non-building objects, leading to inefficient search results and high costs when constructing databases with single-building images.
Innovation Solution
A method and apparatus for constructing a building search database by refining images to detect and extract features of individual buildings using building mask images, calculating area ratios, and selecting keypoints that overlap with building masks to improve search accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If street-view images are used to construct the image database, then image acquisition cost and ease of data collection are improved, but search accuracy deteriorates due to multiple buildings and non-building objects being captured in one scene
Solution Approach 1:
The patent applies segmentation by dividing a single street-view image containing multiple buildings into separate building-specific images. Each building image contains only one building by identifying building regions and extracting them as individual images, thereby resolving the ambiguity of matching query images to specific buildings while maintaining the use of easily acquired street-view data
Solution Approach 2:
The patent extracts building-specific regions from street-view images by detecting building boundaries and creating separate building images. This extraction process removes non-building objects and other buildings from the scene, leaving only the target building for accurate matching and search operations
2Measurement precision
If only single-building images are used to construct the search database, then search accuracy is improved by eliminating ambiguity, but image acquisition cost increases excessively
Solution Approach 1:
The patent performs preliminary processing on readily available street-view images by automatically detecting building regions, segmenting individual buildings, and creating building-specific images before they are used for search. This preliminary action transforms ordinary street-view images into specialized building images without requiring new image acquisitions, thereby maintaining cost-effectiveness while improving search accuracy
Data Source
AI summary
An apparatus and method for searching for a building on the basis of an image and a method of constructing a building search database (DB) for image-based building search. The method includes constructing a building search DB, receiving a query image from a user terminal, detecting a region to which a building belongs in the query image, extracting features of the region detected in the query image, and searching the building search DB for a building matching the extracted features. Therefore, building search performance can be improved.